flan-t5-large-da-multiwoz_500

This model is a fine-tuned version of google/flan-t5-large on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3826
• Accuracy: 37.4297
• Num: 3689
• Gen Len: 16.4142

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 8
• eval_batch_size: 24
• seed: 1799
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Num	Gen Len
1.3527	0.47	200	0.5645	25.0872	3689	12.6606
0.6276	0.93	400	0.4722	31.0261	3689	15.3814
0.539	1.4	600	0.4367	34.1584	3689	15.8056
0.5087	1.86	800	0.4164	35.1677	3689	15.6544
0.4633	2.33	1000	0.4112	34.1615	3689	15.7842
0.463	2.79	1200	0.3961	36.5992	3689	16.4803
0.4437	3.26	1400	0.3895	36.7915	3689	16.5259
0.4328	3.72	1600	0.3874	36.7043	3689	16.2385
0.4189	4.19	1800	0.3826	37.4297	3689	16.4142
0.4239	4.65	2000	0.3804	37.2685	3689	16.2329

Framework versions

• Transformers 4.18.0
• Pytorch 1.10.0+cu111
• Datasets 2.5.1
• Tokenizers 0.12.1